CN116545497A - Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network - Google Patents

Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network Download PDF

Info

Publication number
CN116545497A
CN116545497A CN202310391773.6A CN202310391773A CN116545497A CN 116545497 A CN116545497 A CN 116545497A CN 202310391773 A CN202310391773 A CN 202310391773A CN 116545497 A CN116545497 A CN 116545497A
Authority
CN
China
Prior art keywords
user
satellite
substep
agent
directed graph
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202310391773.6A
Other languages
Chinese (zh)
Other versions
CN116545497B (en
Inventor
李炜
姜治云
杨梦龙
王祥通
韩笑冬
邢川
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan University
Original Assignee
Sichuan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sichuan University filed Critical Sichuan University
Priority to CN202310391773.6A priority Critical patent/CN116545497B/en
Publication of CN116545497A publication Critical patent/CN116545497A/en
Application granted granted Critical
Publication of CN116545497B publication Critical patent/CN116545497B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/14—Relay systems
    • H04B7/15—Active relay systems
    • H04B7/185—Space-based or airborne stations; Stations for satellite systems
    • H04B7/1851—Systems using a satellite or space-based relay
    • H04B7/18519—Operations control, administration or maintenance
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12—Discovery or management of network topologies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method, which comprises the following steps: establishing a user directed graph comprising nodes and directed edges; determining multidimensional attribute weights of the directed edges; establishing a multi-objective optimization model; generating a searching tree root node by utilizing an initial stage algorithm of multi-agent multi-target path planning according to the user directed graph and the multi-target optimization model after the multi-dimensional attribute weight of the directed edge is determined; and searching the generated searching tree root nodes by utilizing a searching stage algorithm of multi-agent multi-target path planning according to the multi-target optimization model to obtain the pareto optimal solution set. According to the invention, the pareto optimal solution set is obtained based on the multi-objective optimization model, namely, the pareto optimal solution set of the multi-user satellite-to-ground link switching strategy, so that the weight of a certain low orbit satellite selected by a user can be determined in the actual implementation process, the switching time delay of the user is reduced, and the switching success rate and the switching efficiency of the user are improved.

Description

Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network
Technical Field
The invention relates to the technical field of satellite communication, in particular to a multi-user satellite-ground link switching method of a multi-target optimized low-orbit satellite network.
Background
To avoid excessive propagation delay of the satellite-to-ground link, the orbit altitude of the low-orbit satellites is typically within 2000 km. This allows the low-orbit satellite nodes to move at very high speeds relative to the ground and provide only a relatively limited coverage to the ground. This aspect illustrates that the satellite-to-ground link between a ground user and a low-orbit satellite can only be maintained for a short period of time. On the other hand, a large number of low-orbit satellites are required to be deployed to realize seamless coverage of the global network, so that users in most areas can be covered by a plurality of satellites at any time. The characteristics of these low-orbit satellite networks provide challenges for the acquisition of the handoff strategies of terrestrial users, and if the handoff satellites cannot be reasonably selected, the quality of satellite-to-ground links is poor and the handoff is frequent, thus affecting the network communication quality of terrestrial users.
For this reason, many efforts have been made to investigate the handover strategy of terrestrial subscribers in low-orbit satellite networks. In the prior art, a user's switching strategy in a period of time is obtained through a directed graph and a graph theory algorithm. However, it requires artificial setting of weights between the user and a certain low-orbit satellite, and the optimization objective under consideration is simpler, without considering the influence between different user handover strategies. In the prior art, when the link quality between the user and the satellite is lower than a threshold value, a target satellite is selected according to three standards of satellite signal quality, satellite load and satellite residual time in sequence. However, it requires a human to determine the threshold and defaults to the priorities of the three decision attributes. In addition, in the prior art, the optimal switching path is obtained in the terminal and satellite switching relation directed graph by adopting a Pareto multi-target genetic algorithm, so that the weight between a user and a certain low-orbit satellite is prevented from being set. However, the optimization index considered by the method is simpler, only the switching strategy of a single user is obtained, and the influence of the switching strategies of other users is ignored.
Disclosure of Invention
Aiming at the defects in the prior art, the multi-user satellite-to-ground link switching method for the multi-target optimized low-orbit satellite network provided by the invention solves the problem that the weight between a user and a certain low-orbit satellite is difficult to determine in the actual implementation process of the multi-user satellite-to-ground link switching method.
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
a multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method comprises the following steps:
s1, establishing a user directed graph comprising nodes and directed edges according to satellite ephemeris;
s2, determining multidimensional attribute weights of the directed edges in the step S1 according to the channel model and the switching relation;
s3, establishing a multi-objective optimization model according to the user directed graph in the step S1;
s4, generating a searching tree root node by utilizing an initial stage algorithm of multi-agent multi-target path planning according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3;
and S5, searching the searching tree root node in the step S4 by utilizing a searching stage algorithm of multi-agent multi-target path planning according to the multi-target optimization model in the step S3, and obtaining the pareto optimal solution set.
Further, step S1 includes the following sub-steps:
s11, dividing a time period into a plurality of time stamps according to network service time required by a user;
s12, determining the distance between each time stamp user divided in the substep S11 and the satellite according to the satellite ephemeris and the user position;
s13, selecting candidate satellites of each time stamp as nodes of the user directed graph according to the distance between the user and the satellite of each time stamp in the substep S12, and generating two empty nodes as a starting point and an end point of the user directed graph;
s14, establishing a directed edge of the user directed graph according to the nodes of the user directed graph in the substep S13;
s15, building the user directed graph according to the nodes of the user directed graph in the substep S13 and the directed edges of the user directed graph in the substep S14.
Further, step S2 includes the following sub-steps:
s21, determining a first dimension attribute weight of the directed edge according to the channel model;
s22, comparing the nodes connected by the directed edges, and determining second dimension attribute weights of the directed edges.
Further, step S21 includes the following sub-steps:
s211, calculating free space propagation loss, which is expressed as:
wherein: FSL (FSL) t In order to be a free-space propagation loss,the distance between the satellite j and the user i when the time stamp t is given, and f is the frequency of the transmission signal;
S212, calculating the user received signal power according to the free space propagation loss in the substep S211, which is expressed as:
wherein:receiving signal power for a user, < >>For the transmitted signal power of satellite j,/>transmit antenna gain for satellite j, +.>The gain of the receiving antenna of the user i is given, and L is the loss caused by the misalignment of the antenna;
s213, obtaining a conditional probability density function relation according to the user received signal power in the substep S212, wherein the conditional probability density function relation is expressed as:
wherein:to take into account the local average signal power +.>In the case of (1) signal strength in the rice fading channel +.>Is a conditional probability density function of->Signal strength, K, provided to user i for satellite j at time t r Is the Lais factor, < >>For the local average signal power, exp is an exponential function based on Euler number e, I 0 Modifying the Bessel function for the first class zero order, < >>For local average signal power +.>S is the shadow diffusion;
s214, substituting the conditional probability density function relation in the substep S213 into the cumulative distribution function of the signal intensity, which is expressed as:
wherein:a cumulative distribution function of signal strength provided to user i by satellite j at time t;
s215, determining a first dimension attribute weight of the directed edge according to the cumulative distribution function of the signal intensity in the substep S214, wherein the first dimension attribute weight is expressed as:
Wherein:directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Index value of satellite closest to user i to the u-th at time stamp t, +.>For the index value of the satellite closest to user i from v when time stamp t+1, y is a random variable subject to standard uniform distribution, +.>Cumulative distribution function of signal strength provided to user i for satellite j at time stamp t>Is an inverse function of (c).
Further, step S3 includes the following sub-steps:
s31, according to the user directed graph in the step S1, a user received signal strength optimization model is established, and the model is expressed as:
wherein:for taking the maximum value of the average received signal strength of the user, < + >>For the average received signal strength of users, N is the total number of users i, f1i is the directed graph G of the intelligent agent i in the users i The first dimension path cost obtained above, agent i corresponds to user i, < >>To sum all paths traversed by agent i, +.>Side +.>Pass judgment value of->Directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Index value of satellite closest to user i to the u-th at time stamp t, +.>An index value of a satellite closest to the user i from the v-th position at time stamp t+1;
s32, building a user switching frequency optimization model according to the user directed graph in the step S1, wherein the user switching frequency optimization model is expressed as follows:
Wherein:for taking the minimum value of the average number of switches of the user, < >>For the average number of switches of the user +.>Directed graph G at user for agent i i The second dimension of the path obtained above, +.>Directed edges for user directed graphsIs a second dimension attribute weight of (2);
s33, establishing a user star selection conflict frequency optimization model according to the user directed graph in the step S1.
Further, step S33 includes the following sub-steps:
s331, determining a user conflict judgment value according to a planning path of the agent, wherein the user conflict judgment value is expressed as:
wherein:user conflict judgment value for judging whether user i conflicts with user j when time stamp t is t time stamp, pi i [t]Path planned for agent i at time stamp t, agent i corresponding to user i, pi j [t]A path planned by an agent j when the time stamp t is given, wherein the agent j corresponds to a user j;
s332, determining the total number of user conflicts according to the user conflict judging value in the substep S331, wherein the total number is expressed as:
wherein: ai, j is the total number of conflicts caused by the agent i to the agent j, and T is the total number of time stamps of the network service time division required by the user;
s333, establishing a user star selection conflict frequency optimization model according to the total number of user conflicts in the substep S332 and the user directed graph in the step S1, wherein the model is expressed as follows:
Wherein:for taking the minimum value of the average number of collisions of users +.>Average number of collisions for the user.
Further, step S4 includes the following sub-steps:
s41, initializing an unsearched node set;
s42, determining a single pareto optimal solution set after optimizing the received signal strength and the switching times by utilizing a single-agent multi-target path planning algorithm according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3;
s43, combining the single pareto optimal solution set after each optimized received signal strength and switching times in the sub-step S42 with the rest single pareto optimal solution sets after all optimized received signal strengths and switching times to obtain a joint path set;
s44, calculating cost vectors of the joint paths in the joint path set in the substep S43 according to the multi-objective optimization model in the step S3;
s45, initializing a constraint set, combining the constraint set with the joint path and the cost vector in the substep S44 to generate a searched tree root node, and adding the searched tree root node to the unsearched node set.
Further, step S5 includes the following sub-steps:
s51, initializing a pareto optimal solution set and a pareto front solution set;
S52, judging whether the unsearched node set is an empty set or not; if yes, outputting a pareto optimal solution set and a pareto front solution set, otherwise, entering a substep S53;
s53, randomly taking out a searching tree root node from the unsearched node set, judging whether a cost vector which can dominate the cost vector of the taken-out searching tree root node exists in the pareto front edge solution set, if so, returning to the substep S52, otherwise, entering the substep S54;
s54, updating the pareto optimal solution set according to the searched tree root nodes extracted in the substep S53;
s55, judging whether the joint paths of the extracted search tree root nodes have conflicts or not; if yes, go to substep S56, otherwise go back to substep S52;
s56, resolving the conflict in the substep S55 by utilizing a conflict resolution mechanism of a CBS algorithm, and generating a path planning constraint;
s57, adding each path planning constraint generated in the substep S56 into the constraint set of the extracted search tree root node to update the constraint set;
s58, executing a single-agent multi-target path planning algorithm on the agents after updating the constraint set in the bisection step S57 to generate a single pareto optimal solution set;
s59, generating new searching tree root nodes according to the single pareto optimal solution set in the substep S58, and judging whether the new searching tree root nodes are put into the unsearched node set;
S510, returning to the substep S52 to search for the next unsearched tree root node.
Further, step S54 includes the following sub-steps:
s541, deleting a cost vector governed by a cost vector of a root node of the extracted search tree in the pareto front solution set and a corresponding joint path in the pareto optimal solution set;
s542, adding the joint paths of the extracted search tree root nodes to the pareto optimal solution set;
s543, adding the cost vector of the extracted root node of the search tree to the pareto front edge solution set.
Further, step S59 includes the following substeps:
s591, creating the joint paths of the extracted search tree root nodes to a set of undetermined joint paths;
s592, determining a pending joint path set according to the single pareto optimal solution set in the substep S58;
s593, judging whether a cost vector exists in the pareto front edge solution set and can dominate the cost vector of the undetermined joint path set in the substep S592; if yes, discarding the pending joint path set and jumping to the sub-step S592, otherwise, entering the sub-step S594;
s594, creating a new searching tree root node according to the updated constraint set in the substep S57, the undetermined joint path set in the substep S592 and the cost vector of the undetermined joint path set in the substep S593, and adding the new searching tree root node to the unsearched node set.
The beneficial effects of the invention are as follows:
(1) The invention directly optimizes the received signal strength of the user, the switching times of the user and the star selecting conflict times of the user by establishing a multi-target optimization model; the user received signal strength is optimized, so that the switching success rate of the user is improved; the optimization of the switching times of the users reduces the accumulated switching time delay of the users; the optimization of the number of the satellite selection conflicts of the users enables the ground users to switch to satellites with light load and sufficient channel number as far as possible; in summary, the multi-objective optimization model is established, so that the switching time delay and the switching success rate of the user are improved, and the switching efficiency of the user is ensured;
(2) The method solves the problem that the weight between a user and a certain low-orbit satellite is difficult to determine in the actual implementation process of the multi-user satellite-to-ground link switching method by establishing a multi-target optimization model;
(3) The invention optimizes the star selecting conflict by utilizing a conflict resolution mechanism in a CBS algorithm, compares the merits of the solutions by utilizing the dominant relationship of the cost vector, and finally obtains the pareto optimal solution set containing a plurality of switching strategies.
Drawings
FIG. 1 is a flow chart of a multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method;
FIG. 2 is a low-orbit satellite network scenario diagram;
FIG. 3 is a directed graph model diagram of a user;
fig. 4 is a diagram of a search process of a multi-agent multi-objective path planning algorithm.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and all the inventions which make use of the inventive concept are protected by the spirit and scope of the present invention as defined and defined in the appended claims to those skilled in the art.
As shown in fig. 1, a multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method includes steps S1-S5:
s1, a user directed graph comprising nodes and directed edges is built according to satellite ephemeris.
In an alternative embodiment of the invention, a directed graph is created for each user in the scene based on satellite ephemeris, representing the handoff relationship of that user between satellites over a period of time. The scenario of the present invention includes M low-orbit satellites and N users, as shown in fig. 2, where ISL is an inter-satellite link (Inter Satellite Link, ISL). The invention uses j to represent index of low orbit satellite, uses i to represent index of user, uses To represent the value set of j byTo represent the set of values of i.
Step S1 comprises the following sub-steps:
s11, dividing the time period into a plurality of time stamps according to the network service time required by the user.
Specifically, the invention divides the whole time period into T time stamps according to the network service time required by the user to form a time stamp setThe index of the timestamp is denoted by t.
S12, determining the distance between each time stamp user divided in the substep S11 and the satellite according to the satellite ephemeris and the user position.
Specifically, the present invention obtains the distance between each time stamped user and each satellite based on satellite ephemeris and user location data.
S13, selecting candidate satellites of each time stamp as nodes of the user directed graph according to the distance between the user and the satellite of each time stamp in the substep S12, and generating two empty nodes as the starting point and the end point of the user directed graph.
Specifically, the invention selects candidate satellites of each time stamp as nodes of the directed graph according to the distance, and the same satellite selected at different time stamps is regarded as different nodes. The present invention generates two empty nodes for each user as the start and end of the user's directed graph, indicating that the user is not connected to any one satellite.
S14, establishing directed edges of the user directed graph according to the nodes of the user directed graph in the substep S13.
Specifically, the invention establishes a directed edge between any two adjacent time stamped nodes that points to the latter time stamped, establishes a directed edge between the starting point and each node of the first time stamped that points to the latter, and establishes a directed edge between each node of the last time stamped that points to the latter and the ending point.
S15, building the user directed graph according to the nodes of the user directed graph in the substep S13 and the directed edges of the user directed graph in the substep S14.
Specifically, the invention uses G i ＝(V i ,E i ) A directed graph representing user i, where V i For the node set of the directed graph, E i For the edge set of the directed graph, usingAnd->Representing a directed graph G i The start point and the end point of (1) have values of null and the rest of nodes are +.>To indicate the index of the satellite closest to user i to k when its value is time stamp t. The user's directed graph model is shown in fig. 3.
S2, determining the multidimensional attribute weight of the directed edge in the step S1 according to the channel model and the switching relation.
In an alternative embodiment of the invention, the invention builds a channel model and sets multidimensional attribute weights for each directed edge in the directed graph using a switching relationship.
Step S2 comprises the following sub-steps:
s21, determining first dimension attribute weight of the directed edge according to the channel model.
In particular, for directed graph G i Directed edge of (a)The invention sets the first dimension weight of the channel model according to the channel model>User i accesses satellite +.>The strength of the signal that can be received,an index value of a satellite closest to the user i to the v-th position at time stamp t+1.
Step S21 includes the following sub-steps:
s211, calculating free space propagation loss, which is expressed as:
wherein: FSL (FSL) t In order to be a free-space propagation loss,the distance between satellite j and user i at time stamp t, f is the transmission signal frequency.
S212, calculating the user received signal power according to the free space propagation loss in the substep S211, which is expressed as:
wherein:receiving signal power for a user, < >>Transmit signal power for satellite j, +.>Transmit antenna gain for satellite j, +.>For the receiving antenna gain of user i, L is the loss due to antenna misalignment.
S213, obtaining a conditional probability density function relation according to the user received signal power in the substep S212, wherein the conditional probability density function relation is expressed as:
wherein:to take into account the local average signal power +.>In the case of (1) signal strength in the rice fading channel +. >Is a conditional probability density function of->Signal strength, K, provided to user i for satellite j at time t r Is the Lais factor, < >>For the local average signal power, exp is an exponential function based on Euler number e, I 0 Modifying the Bessel function for the first class zero order, < >>For local average signal power +.>S is the shadow diffusion.
S214, substituting the conditional probability density function relation in the substep S213 into the cumulative distribution function of the signal intensity, which is expressed as:
wherein:a cumulative distribution function of signal strength provided to user i by satellite j at time t is time stamp t.
S215, determining a first dimension attribute weight of the directed edge according to the cumulative distribution function of the signal intensity in the substep S214, wherein the first dimension attribute weight is expressed as:
wherein:directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Is a time stampIndex value of satellite closest to user i to the u-th at t, < >>For the index value of the satellite closest to user i from v when time stamp t+1, y is a random variable subject to standard uniform distribution, +.>Cumulative distribution function of signal strength provided to user i for satellite j at time stamp t>Is an inverse function of (c).
Specifically, the channel model in the present invention is based on a shadow rice channel that calculates the signal strength provided by satellite j to user i at time stamp t The calculation formula is specifically as follows:
s22, comparing the nodes connected by the directed edges, and determining second dimension attribute weights of the directed edges.
In particular, the invention is implemented by comparing directed edgesNode->And node->If the two nodes are not equal, setting second dimension weight of the directed edge>A value of 1 indicates that a handover has occurred; if two nodesEqual then the second dimension weight of the directed edge is set +.>A value of 0 indicates that no handover has occurred.
S3, establishing a multi-objective optimization model according to the user directed graph in the step S1.
In an optional embodiment of the present invention, the present invention aims at maximizing the average received signal strength of the user, minimizing the average switching times of the user, and minimizing the number of satellite selection collisions of the user, and makes a multi-objective optimization problem based on a directed graph, and the multi-user switching strategy solving problem is converted into a multi-agent path planning problem, where the user i corresponds to the agent i. Based on such thought, the invention establishes a multi-objective optimization model according to the user directed graph in step S1.
Step S3 comprises the following sub-steps:
s31, according to the user directed graph in the step S1, a user received signal strength optimization model is established, and the model is expressed as:
wherein:for taking the maximum value of the average received signal strength of the user, < + > >For the average received signal strength of users, N is the total number of users i, f1i is the directed graph G of the intelligent agent i in the users i The first dimension path cost obtained above, agent i corresponds to user i, < >>To sum all paths traversed by agent i, +.>Side +.>Pass judgment value of->Directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Index value of satellite closest to user i to the u-th at time stamp t, +.>An index value of a satellite closest to the user i to the v-th position at time stamp t+1.
S32, building a user switching frequency optimization model according to the user directed graph in the step S1, wherein the user switching frequency optimization model is expressed as follows:
wherein:for taking the minimum value of the average number of switches of the user, < >>For the average number of switches of the user +.>Directed graph G at user for agent i i The second dimension of the path obtained above, +.>Directed edges for user directed graphsIs included.
S33, establishing a user star selection conflict frequency optimization model according to the user directed graph in the step S1.
Step S33 includes the following sub-steps:
s331, determining a user conflict judgment value according to a planning path of the agent, wherein the user conflict judgment value is expressed as:
wherein:user conflict judgment value for judging whether user i conflicts with user j when time stamp t is t time stamp, pi i [t]Path planned for agent i at time stamp t, agent i corresponding to user i, pi j [t]And (3) a path planned by the agent j when the time is marked t, wherein the agent j corresponds to the user j.
S332, determining the total number of user conflicts according to the user conflict judging value in the substep S331, wherein the total number is expressed as:
wherein: ai, j is the total number of collisions caused by agent i to agent j, T is the total number of time stamps of the network service time divisions required by the user.
S333, establishing a user star selection conflict frequency optimization model according to the total number of user conflicts in the substep S332 and the user directed graph in the step S1, wherein the model is expressed as follows:
wherein:for taking the minimum value of the average number of collisions of users +.>Average number of collisions for the user.
S4, generating a searching tree root node by utilizing an initial stage algorithm of multi-agent multi-target path planning according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3.
In an optional embodiment of the present invention, after the multi-objective optimization model is established, a multi-agent multi-objective path planning algorithm is performed on the directed graphs of all the users to obtain pareto optimal solutions, where each pareto optimal solution corresponds to a switching policy of all the users in a period of time. The search process of the multi-agent multi-objective path planning algorithm is shown in FIG. 4, wherein The joint path generated by the initial stage algorithm for multi-agent multi-target path planning is pi 1 The cost vector is c 1 Search tree root node with constraint set being empty set,>the joint path generated by the initial stage algorithm for multi-agent multi-target path planning is pi 2 The cost vector is c 2 Search tree root node with constraint set being empty set,>the joint path generated by the initial stage algorithm for multi-agent multi-target path planning is pi k The cost vector is c k The constraint set is the searching tree root node of the empty set. The searching process of the multi-agent multi-target path planning algorithm comprises an initial stage algorithm of the multi-agent multi-target path planning and a searching stage algorithm of the multi-agent multi-target path planning, and the multi-agent multi-target path planning method utilizes the multi-agent multi-target path planningAn initial stage algorithm of the multi-objective path planning generates a search tree root node.
Step S4 comprises the following sub-steps:
s41, initializing an unsearched node set.
Specifically, the present invention initializes the unsearched node set OPEN to an empty set.
S42, determining a single pareto optimal solution set after optimizing the received signal strength and the switching times by utilizing a single-agent multi-target path planning algorithm according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3.
Specifically, the invention is based on the user directed graph G after determining the multidimensional attribute weight of the edge i And according to the multi-objective optimization model in step S3, each intelligent agent i executes a single-intelligent-agent multi-objective path search algorithm to obtain a single-user pareto optimal solution set O under the condition of no constraint, with the aim of maximizing the average received signal intensity of the user and minimizing the average switching times of the user i And finally, determining the pareto optimal solution set of all users by utilizing an initial stage algorithm of multi-agent multi-target path planning.
S43, combining the single pareto optimal solution set after each optimized received signal intensity and switching times in the sub step S42 with the rest single pareto optimal solution sets after all optimized received signal intensity and switching times to obtain a joint path set.
Specifically, the invention obtains the single-user pareto optimal solution set O of each intelligent agent in the substep S42 i Combining to obtain a joint path set pi, which is expressed as:
s44, calculating cost vectors of the joint paths in the joint path set in the substep S43 according to the multi-objective optimization model in the step S3.
Specifically, for each joint path pi in the joint path set k ∈Π，According to the multi-objective optimization model in step S3, calculating a cost vector c containing three objective optimization model functions k 。
S45, initializing a constraint set, combining the constraint set with the joint path and the cost vector in the substep S44 to generate a searched tree root node, and adding the searched tree root node to the unsearched node set.
Specifically, the present invention initializes a constraint set Ω k Is an empty set and combines the paths pi k Cost vector c k Constraint set Ω k (an empty set) is combined to generate a search tree root node R k (π k ,c k ,Ω k ) Finally it is added to the unsearched node set OPEN.
And S5, searching the searching tree root node in the step S4 by utilizing a searching stage algorithm of multi-agent multi-target path planning according to the multi-target optimization model in the step S3, and obtaining the pareto optimal solution set.
In an alternative embodiment of the present invention, the search process of the multi-agent multi-objective path planning algorithm is shown in FIG. 4, in whichThe joint path generated by the search phase algorithm for multi-agent multi-target path planning is +.>The cost vector is +.>The constraint set is +.>Is used for searching the root child node of the tree,the joint path generated by the search phase algorithm for multi-agent multi-target path planning isThe cost vector is +. >The constraint set is +.>Is a search tree root child node->The joint path generated by the search phase algorithm for multi-agent multi-target path planning is +.>The cost vector is +.>The constraint set is +.>Is a tree root child node. The searching process of the multi-agent multi-target path planning algorithm comprises an initial stage algorithm of the multi-agent multi-target path planning and a searching stage algorithm of the multi-agent multi-target path planning, and the searching tree root node in the step S4 is searched by utilizing the searching stage algorithm of the multi-agent multi-target path planning to obtain the pareto optimal solution set, namely the joint path in figure 4>The joint path is +.>And joint path->
Step S5 comprises the following sub-steps:
s51, initializing a pareto optimal solution set and a pareto front solution set.
Specifically, the pareto optimal solution set stores the joint paths of all the agents after the multi-agent multi-target path planning algorithm is executed. The pareto front solution set stores the cost vector of each solution, namely each joint path, in the pareto optimal solution set. In the invention, whether the solution found in the algorithm process is the optimal solution of pareto is judged by the dominant relation with the cost vector in the pareto front.
S52, judging whether the unsearched node set is an empty set or not; if yes, outputting the pareto optimal solution set and the pareto front solution set, otherwise, entering a substep S53.
Specifically, the invention judges whether the unsearched node set OPEN is an empty set or not; if yes, go to substep S53, otherwise, output pareto optimal solution set P and pareto front solution set P 1 The pareto optimal solution set is obtained by the method, namely the pareto optimal solution set of the multi-user satellite-to-ground link switching strategy.
S53, randomly taking out a searching tree root node from the unsearched node set, judging whether a cost vector exists in the pareto front solution set and can dominate the cost vector of the taken-out searching tree root node, if yes, returning to the substep S52, otherwise, entering the substep S54.
Specifically, the invention takes a searching tree root node R from an unsearched node set OPEN k (π k ,c k ,Ω k ) And judge the solution set P at the pareto front 1 Whether or not there is a cost vector can dominate the cost vector c k If yes, go back to substep S52, otherwise go to substep S54.
And S54, updating the pareto optimal solution set according to the searched tree root nodes extracted in the substep S53.
Step S54 includes the following sub-steps:
s541, deleting the cost vector governed by the cost vector of the root node of the extracted search tree in the pareto front solution set and the corresponding joint path in the pareto optimal solution set.
In particular, the present invention solves the pareto front solution set P 1 Tree root node R of the medium-searched tree k (π k ,c k ,Ω k ) Cost vector c of (a) k Those cost vectors that are dominant and the corresponding joint path deletions in the pareto optimal solution set P.
S542, adding the joint paths of the extracted search tree root nodes to the pareto optimal solution set.
Specifically, the invention searches the root node R of the tree to be fetched k (π k ,c k ,Ω k ) Is of the joint path pi of (a) k Added to pareto optimal solution set P.
S543, adding the cost vector of the extracted root node of the search tree to the pareto front edge solution set.
Specifically, the invention searches the root node R of the tree to be fetched k (π k ,c k ,Ω k ) Cost vector c of (a) k Added to pareto front solution set P 1 Is a kind of medium.
S55, judging whether the joint paths of the extracted search tree root nodes have conflicts or not; if yes, go to substep S56, otherwise go back to substep S52.
Specifically, the invention judges the extracted searching tree root node R k (π k ,c k ,Ω k ) Is of the joint path pi of (a) k Whether there is a conflict. In the actual operation process, the invention looks at the joint path pi in the directed graph of all users k If the paths of the agents under the policy generate node conflict, that is, if a plurality of agents pass through the same node at the same time, the method goes to the substep S55 if yes, otherwise goes back to the substep S52.
S56, resolving the conflict in the substep S55 by utilizing a conflict resolution mechanism of the CBS algorithm, and generating a path planning constraint.
Specifically, the invention utilizes Conflict Based Search (CBS, conflict-based search algorithm) algorithm to solve the search tree root node R k (π k ,c k ,Ω k ) Is of the joint path pi of (a) k Generates a path plan constraint Ω. Under the conflict resolution mechanism of Conflict Based Search algorithm, if agent A and agent B conflict at a certain node, two constraints are generated, one constraint agent A cannot pass through the conflict point, and the other constraint agent A cannot pass through the conflict pointBeam agent B cannot pass through the conflict point. In the invention, only one constraint is generated according to the user priority, so that the generated constraint is reduced, the search time is reduced, and excessive search tree nodes are avoided. If the path planning constraint of the agent i is generatedThe path planning constraint is expressed as that the path of agent i cannot pass through the conflicting node +.>
S57, adding each path planning constraint generated in the substep S56 into the constraint set of the extracted search tree root node to update the constraint set.
In particular, the present invention relates to each of the path planning constraints ω i Constraint omega is planned for each path i And the extracted searching tree root node R k (π k ,c k ,Ω k ) Path planning constraint Ω of (2) k Taking the union set to generate a new constraint set omega l To update the constraint set.
And S58, executing a single-agent multi-target path planning algorithm on the agents after updating the constraint set in the bisection step S57, and generating a single pareto optimal solution set.
Specifically, a new set of constraint sets Ω is generated l Thereafter, the present invention is constrained at user iThen, the intelligent agent i runs a single intelligent agent multi-target path planning algorithm to obtain an individual pareto optimal solution set +.>For each +.>
S59, generating new searching tree root nodes according to the single pareto optimal solution set in the substep S58, and judging whether the new searching tree root nodes are put into the unsearched node set.
Step S59 includes the following sub-steps:
s591, creating the joint paths of the extracted search tree root nodes to a set of undetermined joint paths.
Specifically, the invention copies the extracted searching tree root node R k (π k ,c k ,Ω k ) Is of the joint path pi of (a) k To pending joint path set pi l Is a kind of medium.
The present invention obtains a single pareto optimal solution set for an agent in step S58, where a plurality of solutions, i.e., a plurality of joint paths, are included. For each joint path, the present invention replicates joint path pi k For joint path pi l 。
If the copy operation is not performed, directly pair the joint path pi k If updated, will have an effect on the next cycle, as in step S57, a constraint on agent i and a constraint on agent j are generated. The present invention first performs step S57 and step S58 on the constraint of agent i, and replaces the joint path pi with the solution of agent i generated in step S58 in step S59 k Solution of agent i. On the next cycle, we perform steps S57 and S58 on the constraints of agent j and replace the joint path pi with the solution of agent j generated in step S58 in step S59 k Solution of agent j, at this time, joint path pi k Since the solution of agent i in the previous cycle replaced the joint path pi k Solution of agent i results in a joint path pi k A change occurs. The present invention performs a copy operation at this sub-step.
S592, determining a pending joint path set according to the single pareto optimal solution set in the substep S58.
Specifically, the present invention uses each joint path in the single pareto optimal solution set in step S58Substitution of pending joint Path set pi in substep S591 l Is +. >Final determination of pending joint Path set pi l 。
The invention aims at generating a new searching tree root child node containing a constraint set in the substep. The solution generated in step S58 of the present invention satisfies the path constraints, so that the agent does not collide with other agents under these solutions, i.e., under these joint paths. The solution of the agent is replaced by a joint path pi k Replicated joint path pi l Corresponding to the solution of the agent, then at this point in the joint path pi l And joint path pi k In contrast, the joint path pi l The agent can not conflict any more under the strategy, and the optimization of the conflict is realized.
S593, judging whether a cost vector exists in the pareto front edge solution set and can dominate the cost vector of the undetermined joint path set in the substep S592; if so, the pending joint path set is discarded and jumps to a sub-step S592, otherwise, the process proceeds to a sub-step S594.
Specifically, the method calculates the pending joint path set pi according to the multi-objective optimization model l Cost vector c of (a) l Then judging the pareto front solution set P 1 Whether or not there is a cost vector that can dominate the cost vector c of the set of pending joint paths in substep S592 l The method comprises the steps of carrying out a first treatment on the surface of the If so, the pending joint path set is discarded and jumps to a sub-step S592, otherwise, the process proceeds to a sub-step S594.
S594, creating a new searching tree root node according to the updated constraint set in the substep S57, the undetermined joint path set in the substep S592 and the cost vector of the undetermined joint path set in the substep S593, and adding the new searching tree root node to the unsearched node set.
Specifically, the invention is based on the updated constraint set omega l Pending joint path set pi l And a cost vector c of the pending joint path set l Creating a new search tree rootNode R l (π l ,c l ,Ω l ) And adds it to the unsearched node set OPEN.
Those of ordinary skill in the art will recognize that the embodiments described herein are for the purpose of aiding the reader in understanding the principles of the present invention and should be understood that the scope of the invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific modifications and combinations from the teachings of the present disclosure without departing from the spirit thereof, and such modifications and combinations remain within the scope of the present disclosure.

Claims (10)

1. The multi-user satellite-ground link switching method for the multi-target optimized low-orbit satellite network is characterized by comprising the following steps of:
s1, establishing a user directed graph comprising nodes and directed edges according to satellite ephemeris;
S2, determining multidimensional attribute weights of the directed edges in the step S1 according to the channel model and the switching relation;
s3, establishing a multi-objective optimization model according to the user directed graph in the step S1;
s4, generating a searching tree root node by utilizing an initial stage algorithm of multi-agent multi-target path planning according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3;
and S5, searching the searching tree root node in the step S4 by utilizing a searching stage algorithm of multi-agent multi-target path planning according to the multi-target optimization model in the step S3, and obtaining the pareto optimal solution set.
2. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 1, wherein the step S1 comprises the following sub-steps:
s11, dividing a time period into a plurality of time stamps according to network service time required by a user;
s12, determining the distance between each time stamp user divided in the substep S11 and the satellite according to the satellite ephemeris and the user position;
s13, selecting candidate satellites of each time stamp as nodes of the user directed graph according to the distance between the user and the satellite of each time stamp in the substep S12, and generating two empty nodes as a starting point and an end point of the user directed graph;
S14, establishing a directed edge of the user directed graph according to the nodes of the user directed graph in the substep S13;
s15, building the user directed graph according to the nodes of the user directed graph in the substep S13 and the directed edges of the user directed graph in the substep S14.
3. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 1, wherein the step S2 comprises the following sub-steps:
s21, determining a first dimension attribute weight of the directed edge according to the channel model;
s22, comparing the nodes connected by the directed edges, and determining second dimension attribute weights of the directed edges.
4. A multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 3, wherein step S21 comprises the following sub-steps:
s211, calculating free space propagation loss, which is expressed as:
wherein: FSL (FSL) t In order to be a free-space propagation loss,the distance between the satellite j and the user i when the time stamp t is given, and f is the frequency of the transmission signal;
s212, calculating the user received signal power according to the free space propagation loss in the substep S211, which is expressed as:
wherein:receiving signal power for a user, < >>Transmit signal power for satellite j, +. >Transmit antenna gain for satellite j, +.>The gain of the receiving antenna of the user i is given, and L is the loss caused by the misalignment of the antenna;
s213, obtaining a conditional probability density function relation according to the user received signal power in the substep S212, wherein the conditional probability density function relation is expressed as:
wherein:to take into account the local average signal power +.>In the case of (1) signal strength in the rice fading channel +.>Is a conditional probability density function of->Signal strength, K, provided to user i for satellite j at time t r Is the Lais factor, < >>For the local average signal power, exp is an exponential function based on Euler number e, I 0 Modifying the Bessel function for the first class zero order, < >>For local average signal power +.>S is the shadow diffusion;
s214, substituting the conditional probability density function relation in the substep S213 into the cumulative distribution function of the signal intensity, which is expressed as:
wherein:a cumulative distribution function of signal strength provided to user i by satellite j at time t;
s215, determining a first dimension attribute weight of the directed edge according to the cumulative distribution function of the signal intensity in the substep S214, wherein the first dimension attribute weight is expressed as:
wherein:directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Index value of satellite closest to user i to the u-th at time stamp t, +. >For the index value of the satellite closest to user i from v when time stamp t+1, y is a random variable subject to standard uniform distribution, +.>Cumulative distribution function of signal strength provided to user i for satellite j at time stamp t>Is an inverse function of (c).
5. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 1, wherein the step S3 comprises the following sub-steps:
s31, according to the user directed graph in the step S1, a user received signal strength optimization model is established, and the model is expressed as:
wherein:for taking the maximum value of the average received signal strength of the user, < + >>For the average received signal strength of the users, N is the total number of users i, f 1 i Directed graph G at user for agent i i The first dimension path cost obtained above, agent i corresponds to user i, < >>To sum all paths traversed by agent i, +.>Is on edge for agent iPass judgment value of->Directed edges of the directed graph for the user +.>Is/are set by the first dimension attribute weight->Index value of satellite closest to user i to the u-th at time stamp t, +.>An index value of a satellite closest to the user i from the v-th position at time stamp t+1;
s32, building a user switching frequency optimization model according to the user directed graph in the step S1, wherein the user switching frequency optimization model is expressed as follows:
Wherein:for taking the minimum value of the average number of switches of the user, < >>For the average number of switches of the user +.>Directed graph G at user for agent i i The second dimension of the path obtained above, +.>Directed edges for user directed graphsIs a second dimension attribute weight of (2);
s33, establishing a user star selection conflict frequency optimization model according to the user directed graph in the step S1.
6. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 5, wherein step S33 comprises the following sub-steps:
s331, determining a user conflict judgment value according to a planning path of the agent, wherein the user conflict judgment value is expressed as:
wherein:user conflict judgment value for judging whether user i conflicts with user j when time stamp t is t time stamp, pi i [t]Path planned for agent i at time stamp t, agent i corresponding to user i, pi j [t]Time stamp t time intelligenceA path planned by the energy j, wherein the intelligent agent j corresponds to the user j;
s332, determining the total number of user conflicts according to the user conflict judging value in the substep S331, wherein the total number is expressed as:
wherein: a, a i,j The total number of conflicts caused by the agent i to the agent j is the total number of time stamps divided by the network service time required by the user;
S333, establishing a user star selection conflict frequency optimization model according to the total number of user conflicts in the substep S332 and the user directed graph in the step S1, wherein the model is expressed as follows:
wherein:for taking the minimum value of the average number of collisions of users +.>Average number of collisions for the user.
7. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 1, wherein the step S4 comprises the following sub-steps:
s41, initializing an unsearched node set;
s42, determining a single pareto optimal solution set after optimizing the received signal strength and the switching times by utilizing a single-agent multi-target path planning algorithm according to the user directed graph with the multi-dimensional attribute weight of the directed edge determined in the step S2 and the multi-target optimization model in the step S3;
s43, combining the single pareto optimal solution set after each optimized received signal strength and switching times in the sub-step S42 with the rest single pareto optimal solution sets after all optimized received signal strengths and switching times to obtain a joint path set;
s44, calculating cost vectors of the joint paths in the joint path set in the substep S43 according to the multi-objective optimization model in the step S3;
S45, initializing a constraint set, combining the constraint set with the joint path and the cost vector in the substep S44 to generate a searched tree root node, and adding the searched tree root node to the unsearched node set.
8. The multi-objective optimized low-orbit satellite network multi-user satellite-ground link switching method according to claim 7, wherein the step S5 comprises the following sub-steps:
s51, initializing a pareto optimal solution set and a pareto front solution set;
s52, judging whether the unsearched node set is an empty set or not; if yes, outputting a pareto optimal solution set and a pareto front solution set, otherwise, entering a substep S53;
s53, randomly taking out a searching tree root node from the unsearched node set, judging whether a cost vector which can dominate the cost vector of the taken-out searching tree root node exists in the pareto front edge solution set, if so, returning to the substep S52, otherwise, entering the substep S54;
s54, updating the pareto optimal solution set according to the searched tree root nodes extracted in the substep S53;
s55, judging whether the joint paths of the extracted search tree root nodes have conflicts or not; if yes, go to substep S56, otherwise go back to substep S52;
s56, resolving the conflict in the substep S55 by utilizing a conflict resolution mechanism of a CBS algorithm, and generating a path planning constraint;
S57, adding each path planning constraint generated in the substep S56 into the constraint set of the extracted search tree root node to update the constraint set;
s58, executing a single-agent multi-target path planning algorithm on the agents after updating the constraint set in the bisection step S57 to generate a single pareto optimal solution set;
s59, generating new searching tree root nodes according to the single pareto optimal solution set in the substep S58, and judging whether the new searching tree root nodes are put into the unsearched node set;
s510, returning to the substep S52 to search for the next unsearched tree root node.
9. The method for multi-user satellite-to-ground link switching for a multi-target optimized low-orbit satellite network according to claim 8, wherein step S54 comprises the following sub-steps:
s541, deleting a cost vector governed by a cost vector of a root node of the extracted search tree in the pareto front solution set and a corresponding joint path in the pareto optimal solution set;
s542, adding the joint paths of the extracted search tree root nodes to the pareto optimal solution set;
s543, adding the cost vector of the extracted root node of the search tree to the pareto front edge solution set.
10. The multi-objective optimized low-orbit satellite network multi-user satellite-to-ground link switching method according to claim 8, wherein step S59 comprises the following sub-steps:
S591, creating the joint paths of the extracted search tree root nodes to a set of undetermined joint paths;
s592, determining a pending joint path set according to the single pareto optimal solution set in the substep S58;
s593, judging whether a cost vector exists in the pareto front edge solution set and can dominate the cost vector of the undetermined joint path set in the substep S592; if yes, discarding the pending joint path set and jumping to the sub-step S592, otherwise, entering the sub-step S594;
s594, creating a new searching tree root node according to the updated constraint set in the substep S57, the undetermined joint path set in the substep S592 and the cost vector of the undetermined joint path set in the substep S593, and adding the new searching tree root node to the unsearched node set.
CN202310391773.6A 2023-04-12 2023-04-12 Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network Active CN116545497B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310391773.6A CN116545497B (en) 2023-04-12 2023-04-12 Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310391773.6A CN116545497B (en) 2023-04-12 2023-04-12 Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network

Publications (2)

Publication Number Publication Date
CN116545497A true CN116545497A (en) 2023-08-04
CN116545497B CN116545497B (en) 2024-01-30

Family

ID=87446044

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310391773.6A Active CN116545497B (en) 2023-04-12 2023-04-12 Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network

Country Status (1)

Country Link
CN (1) CN116545497B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107291090A (en) * 2017-05-12 2017-10-24 北京空间飞行器总体设计部 A kind of continuous imaging control method optimized based on critical path
CN112491619A (en) * 2020-11-25 2021-03-12 东北大学 Self-adaptive distribution technology for service customized network resources based on SDN
CN112558474A (en) * 2020-11-30 2021-03-26 重庆邮电大学 Low-orbit satellite communication line switching control method based on multi-target genetic algorithm
WO2022043512A1 (en) * 2020-08-28 2022-03-03 UMNAI Limited Behavior modeling, verification, and autonomous actions and triggers of ml and ai systems

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107291090A (en) * 2017-05-12 2017-10-24 北京空间飞行器总体设计部 A kind of continuous imaging control method optimized based on critical path
WO2022043512A1 (en) * 2020-08-28 2022-03-03 UMNAI Limited Behavior modeling, verification, and autonomous actions and triggers of ml and ai systems
CN112491619A (en) * 2020-11-25 2021-03-12 东北大学 Self-adaptive distribution technology for service customized network resources based on SDN
CN112558474A (en) * 2020-11-30 2021-03-26 重庆邮电大学 Low-orbit satellite communication line switching control method based on multi-target genetic algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
WENLIANG LIN等: ""A Novel Method to Determine the Handover Threshold Based on Reconfigurable Factor Graph for LEO Satellite Internet Network"", 《IEEE ACCESS》 *
ZHAOFENG WU等: ""A Graph-Based Satellite Handover Framework for LEO Satellite Communication Networks"", 《 IEEE COMMUNICATIONS LETTERS》 *

Also Published As

Publication number Publication date
CN116545497B (en) 2024-01-30

Similar Documents

Publication Publication Date Title
CN113038387B (en) Switching judgment method based on Q learning in low orbit satellite network
CN109889255B (en) Satellite network reconstruction method based on improved bee colony algorithm
Azzimonti et al. Active vs transfer learning approaches for QoT estimation with small training datasets
EP2369880A1 (en) Method and device for synchronizing nodes of a mesh wireless network
US20220104034A1 (en) Method of association of user equipment in a cellular network according to a transferable association policy
CN112804726A (en) Multi-agent reinforcement learning routing algorithm based on geographical position
CN116545497B (en) Multi-objective optimized multi-user satellite-ground link switching method for low-orbit satellite network
Anh et al. A deep reinforcement learning approach for backscatter-assisted relay communications
Perlaza et al. On the base station selection and base station sharing in self-configuring networks
CN105764145B (en) The method of power optimization distribution in the double relay systems of increment based on error sign ratio
CN112803984B (en) Method for switching between satellites in space-earth integrated communication network
Loevenich et al. A bayesian inference model for dynamic neighbor discovery in tactical networks
Ma et al. Socially aware distributed caching in device-to-device communication networks
CN115150335B (en) Optimal flow segmentation method and system based on deep reinforcement learning
Zhang et al. Multi-objective satellite selection strategy based on entropy
HasanzadeZonuzy et al. Reinforcement learning for multi-hop scheduling and routing of real-time flows
Huang et al. DTN routing algorithm based on service probability and limited copy for satellite networks
Wong et al. Adaptive phase synchronisation algorithm for collaborative beamforming in wireless sensor networks
Tao et al. An Efficient Multipath Routing Algorithm Based on LEO Satellite Networks
Kharbash et al. All-terminal network reliability optimization in fading environment via cross entropy method
CN115361055B (en) Inter-satellite switching method of satellite communication system based on user group
Hong et al. Parallel Subnetwork Routing Algorithm for Inter-Satellite Optical Communication
Ortega Finding repeater placement for P2P wireless links with NLOS in extremely mountainous regions
CN117978250A (en) Low orbit satellite switching strategy solving method based on dynamic programming
Kalathil et al. Multi-player multi-armed bandits: Decentralized learning with IID rewards

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant